Semi-Automatic Construction of Ontologies from Text
The Master's Thesis deals with semi-automatic construction of ontologies from text. While the core of the thesis was to develop an integrated system for ontology population with instances extracted from text, it also discusses and analyzes two major existing approaches in this area. The system is based on supervised learning and therefore learns extraction rules from annotated text and then applies those rules on new documents for the extraction. The important part of the entire cycle of ontology population is a user who accepts, rejects or modifies new extractions and suggested instances to be populated. An analysis of the possibility of automatically creation of new classes is discussed in turn.